Hi there, I'm Dmytro Melnyk! 👋

Backend Systems & Platform Engineer | Distributed Systems | Kubernetes

Building resilient, high-throughput distributed systems and deep-system observability tools.
Passionate about Go internals, Linux Kernel (eBPF), Consensus Protocols (Raft), and Cloud-Native orchestration.

I spend my free time building infrastructure tooling, kernel-level tracers, and cloud-native control planes from scratch:

• 🐝 eBPF TCP Connection Scanner — A zero-instrumentation Linux network observability tool. Uses eBPF (kprobes) written in C to hook into the kernel's TCP stack (tcp_sendmsg), tracking active connections, PIDs, and TX byte throughput in real-time.
• 🚦 Kubernetes Gateway API Controller — A custom ingress proxy implementing the modern Gateway API (HTTPRoute). Features a controller-runtime control plane watching EndpointSlices, and an embedded reverse proxy data plane with an atomic Round-Robin load balancer that routes traffic directly to Pod IPs.
• 🌲 LSM-Tree Storage Engine — A high-performance, embedded key-value storage engine engineered from scratch in Go. Features crash-safe WAL (CRC32), lock-free SkipList MemTables, Bloom Filters, and K-way merge iterators (Min-Heap) for seamless memory-to-disk range scans.
• 📦 Distributed KV Storage — A custom, high-performance distributed key-value database written in Go. Features a from-scratch Raft consensus implementation (Leader Election, Log Replication, Snapshots, Read Index) and a Bitcask-style storage engine.
• 🔄 Distributed CRDT Node (OR-Set) — A masterless, eventually consistent data structure in Go. Implements an Observed-Remove Set with Add-Wins semantics. Uses Vector Clocks (DotContext) for causal tracking and a background P2P Gossip protocol for delta-state synchronization and conflict-free convergence.
• ☸️ KV Kubernetes Operator — A custom K8s Operator built with Kubebuilder to automate the deployment, Raft bootstrapping, and lifecycle management of the kv-storage cluster via CRDs and StatefulSets.
• 📈 Binance Market Data Ingestor — High-frequency crypto data pipeline. Streams WebSocket ticks, stores raw order books in ClickHouse, and calculates 1-minute OHLCV candles on-the-fly using SummingMergeTree & Materialized Views.
• 🛴 Fleet Telemetry System — High-throughput IoT tracking backend. Uses a custom 26-byte binary protocol, NATS JetStream, and Uber H3 geospatial indexing to track 10k+ concurrent connections.

My professional and open-source work primarily revolves around:

• Cloud-Native Orchestration: Extending Kubernetes via custom Controllers, Operators (Kubebuilder), and modern routing standards (Gateway API).
• Linux Kernel Observability: Writing custom eBPF programs in C, safely interacting with kernel space via BPF Maps, and building lockless IPC with Go user-space daemons.
• Distributed Systems: Implementing consensus algorithms (Raft), handling split-brain scenarios, and ensuring linearizability.
• High-Frequency Processing: Building zero-allocation Go pipelines, custom network proxies, and real-time event streaming (NATS JetStream).
• Storage & OLAP: Engineering custom storage layers (LSM-Trees/Bitcask) and optimizing real-time aggregations in ClickHouse.